合肥工业大学学报(自然科学版)
閤肥工業大學學報(自然科學版)
합비공업대학학보(자연과학판)
Journal of Hefei University of Technology (Natural Science)
2015年
10期
1358-1362
,共5页
陈强云%郭铭铭%窦建华
陳彊雲%郭銘銘%竇建華
진강운%곽명명%두건화
图像去噪%非局部均值%方向结构信息%图像冗余信息%邻域相似性%相似性度量
圖像去譟%非跼部均值%方嚮結構信息%圖像冗餘信息%鄰域相似性%相似性度量
도상거조%비국부균치%방향결구신식%도상용여신식%린역상사성%상사성도량
image denoising%non-local mean value%directional structural information%image redun-dant information%neighborhood similarity%similarity measurement
文章针对传统的非局部均值算法只关注图像几何结构信息而忽略图像的方向结构信息,提出了一种基于增强邻域结构方向信息的非局部均值算法。通过增强邻域结构方向信息能够更加合理地描述邻域间的相似度,使相似度高的邻域获得更高的权重,相似性度量也具有较强的鲁棒性。实验结果表明,算法能够取得很好的去噪效果,同时能够保留图像的边缘结构信息,特别在结构比较复杂的区域体现得更加明显。
文章針對傳統的非跼部均值算法隻關註圖像幾何結構信息而忽略圖像的方嚮結構信息,提齣瞭一種基于增彊鄰域結構方嚮信息的非跼部均值算法。通過增彊鄰域結構方嚮信息能夠更加閤理地描述鄰域間的相似度,使相似度高的鄰域穫得更高的權重,相似性度量也具有較彊的魯棒性。實驗結果錶明,算法能夠取得很好的去譟效果,同時能夠保留圖像的邊緣結構信息,特彆在結構比較複雜的區域體現得更加明顯。
문장침대전통적비국부균치산법지관주도상궤하결구신식이홀략도상적방향결구신식,제출료일충기우증강린역결구방향신식적비국부균치산법。통과증강린역결구방향신식능구경가합리지묘술린역간적상사도,사상사도고적린역획득경고적권중,상사성도량야구유교강적로봉성。실험결과표명,산법능구취득흔호적거조효과,동시능구보류도상적변연결구신식,특별재결구비교복잡적구역체현득경가명현。
As traditional non‐local mean value algorithm only concerns about geometric structural infor‐mation of the image w hile ignores the directional structural information ,a new non‐local mean value algorithm based on directional information of enhanced neighboring structure is proposed .The simi‐larity between the neighbor fields can be described more reasonably by directional information of en‐hanced neighboring structure .It makes more similar neighborhoods obtain higher weights and the measurement of similarity is more robust .The experimental results show that the new algorithm a‐chieves good denoising effect and reserves marginal information simultaneously .The result is much better especially in the area w here structure is more complicated .